65

A REVISION OF AUSTMLOPITHECINE BODY SIZES

Grover S. Krantz Anthropology Department Washington State University Pullman, Washington 99L63, U.S.A. Received February 18, L977

ABSTMCT: When AustralopiEhecus dentiEions are divided into africanus and robustus on morphological grounds alone it is found that both species are found in each of the major South African sites. There is a greaL sexual dimor- phisrn in both species with a relative scarcity of the big urale africanus and the small female robustus. Skulls and jaws of young male africanus and female robustus have previously been misidentified as early genus Homg. With site allocation no longer valid, a ner^rmethod of identifying postcranial remains shows that africanus averages much larger than robustus. With new ratios of tooth fo body size the greater degree of morphological molarization in robustus now makes sense. J

Introduction

Australopithecus remains are generally thought of as two species; A. elg- ttgracilet' t'robusttt canus is the form, and A. robustus is the form. Some think this distinction merely reflects differences beLween sexes, races, and indivi- dual variations within a single species. Ot.hers raise the distinction to generic level with Paranghlgpge being the "robust't form, and Australopithecus continu- ing as tfre \ra"ifettorm which is sometimes even ittcl.rded in our own genus as Homo africanus. These various opinions are well known in the literature. The taxonomic leve1 of distinction between the trvo forms i-s not at issue here, but rather the most basic assumptions on which the two forms are based. Before anything meaningful can be decided about two kj-nds of australopithecines it is necessary to know which specimens belong in each category and how they differ from each other. The following discussj-on will center mainly on the South African specimens and extend from Lhere as pertinent. To most readers this is already well knownl robustus is the large form found at Swartkrans and Kromdraai, while africanus is the small form found at Taung, Sterkfontein, and Makapansgat. Other differences often discussed include molarization, cranial architecture, crest development, and locomotor adaptations. There is much disagreement about the nature, significance, or even the renli gy of these other differences, though everyone seems to agree they can be distinguished by sLze and by site. But apparently we have all been misled on both pointsl Two types of australopithecines can be distinguished in terms of dentitions, but it will be shown here that both of these occur mixed together in the vari-- ous sites. Since this would invalidate postcranial assignments based on the sites of recovery, other methods must be used for such assignments. The other methods developed here show that africanus is actually larger than robustus. This conclusion, if demonstrated, serves to clarify some problems, and may somewhat affect the phylogenies that are drawn up relating these forms to each other and to other hominids. It ls somewhat of a mystery to me how the australopithecine remains came Lo be classified according to site. Nowhere in the literature can one find a a:kt( Evol. Theory 2265-94 (May, L977) The editors thank two anonymous referees for help in evaluating this paper. @L977, The UniversitY of Chicago. 66 G. S. KRANTZ clear statement of how it is known that all specimens from a given South African cave deposit are of the same type. Somehave wondered about such site exclu- siveness, but that is about as far as it has gone. This arbitrary categorization has been uni-versally accepted by workers who have made comparisons between what they consider to be samples of the two forms. In all normal paleontology any pooling of specimens into a sample rests on the identification of every single individual as belonging to that taxon. Even in the australopithecine sites, other ani-mals are handled properly. If a few hyenas in one site are identified as belonging to a given species it is noE claimed that all hyenid remains from that site autornatically belong to that same species. It has not been expl-ained why the hominid remains were pooled without regard to demonstrating the affinities of each individual specimen. Some question has recently been raised that Makapansgat may contain speci- mens of both forms, or that it may represent a sample drawn from the point of time where they were only beginning to differentiate from each other (Tobias, L973). This idea is supported by observations on the fossils themselves (Wal- lace, L973, and Aguirre, 1970) and also by the strong possibility that Maka- pansgat may in fact be the earliest of all theaustralopithecinesites (We11sr1969). The universal- acceptance that robustus is larger in body size than africanus follows directly from this al-locatioi=byGite. Some hominid postcraniaT=66G- of fairly large size were found at Swartkrans and some very much smaller bones come from Sterkfontein. Since these constitute most of the hominid bones, and they correspond to the size contrast of rnost of the skull-s from these two sites, the body size contrast is easil-y postul-ated. The bones that donrt fit this picture, smaLl ones from Swartkrans and large ones from Sterkfontein, are either ignored or referred to other taxa entirely. In no case can cranial and post- cranial remains be related to the same individual-; mere physicaL proximity is not sufficient in these random bone accumul-ations. From here it is a simple matter to cl-assify any smal-l australopithecine bone as africanus, and every large one as robustus, wherever they are found. But if site alLoeation is not aLways correct, then the size distinction night not always be correct, and the ramificatlons get interesting. A11 of the evidence and reasoning behind these assignments needs to be re-examined with some care, and a modest beginning on this will be made here.

Two Dental TYPes

The usual pracLice has been to equate degree of molarizatLon with size and site. The more molarized dentitions go with larger body sizes and come from "robustn sites. In order to test the reality of this assumed correlation it is first necessary to divide the known dentitions into two morphologically distinct types without regard to their place of origin. This can be done rather easily for many speci-mens with the published measurements and descriptions. The length and breadth of each measurable tooth are here urultiplied together to give approximatlons of the cror^rn surface area. These are exaggerated because the teeth are not rectangular, but the exagerat.ionis consistent in all specimens. These surface areas are then arranged in the form of a bar graph for each jaw, beginning with the first incisor on the left and running to the third molar. The line connecting the tops of these area measurements constitutes what rnight be called a dental profile. So far this procedure is not greatly different from that used by Robinson in manypublications to illustrate the sequence of sizes of australopitheeine teeth. One minor difference is that instead of area he used dental module, the mean of length and breadth, which relatively exaggerates those teeth that are more nearly square. 67

AUSTMLOPITHECINEBODY SIZES

A major difference is that Robinson, like all others, made his definitive comparisons between the dental profiles of pooled samples. Means and ranges of "robust" specimens were conLrasted with those of 'tgracileil specimens on the assumption that all hominid teeth from each site belonged to a particular taxon. My procedure here has been to plot the more complete dental profiles of sepa- rate individuals and to describe the types that occur. Then all possible speci- mens are compared individually and classified according to which type thei-r dental profiles most closely resemble. Obviously this cannot be complete as many specimens include just one or a few teeth and show no clear pattern. If this procedure had put all el-assifiable individuals from Swartkrans and Kromdraai in one category and all those from Sterkfontein and Makapansgat in the other, it might then have been a good guess (and no more) that all other individuals would likewise belong to the type traditionally associated with their sites. Actually the first sorting, by dental type, failed to support the accept,ed site allocati.ons. In contructing dental profiles it soon became obvious that M 3 was an erratic tooth. Its area often bore l-ittle relation to that of the preceding moIar. In many cases Lwo or more individuals with otherwise si-milar profiles differed greatly in the size of their last molars. Also there were many in- stances where the right and left ttwisdomtt teeth of the same individual were markedly different. Accordingly, this tooth was left out of any serious con- sideration ln assignments to types. The first incisors were also found to be more confusing than helpful, but in a different r,ray. In the same individuals (and those determined to be of the same type) the upper and lower dental profiles proved to be remarkably similar except for these first incisors. The tooth is markedly larger in the upper jaw. This is quite natural as its great mesiodist,al J-ength ensures that all sub- sequent upper teeth overl-ap their lower counterparts distally by half a cusp 1-ength. This results in cusp alternations which tend to keep newly erupted teeth in their proper occlusal reLationships until their roots are firmly es- tabllshed. If this first incisor i"s ignored, the terminology used to describe the shapes of the dental profiles is exactly the same for both upper and lower jaws. The various dental profiles are also compared with those of genus Homo as drawn from data conpiled by I'Iolpoff (197f). These are pooled measurements of H. erectus and H. sapiens, shown separately, but the large numbers are suffi- Aient to overriEe any in.orrect individual assignments. I did make one change in substituting neandertal means for those of erectus wherever the former were larger. This was done partly because I prefer to class neandertals in that species, and partly because the resulting profile parallels that of sapiens more closely than do the accepted erectus by themselves. Australopithecine measurements are also taken from Wolpoff (f971), corrected in some cases from Robinson (1956), and some individuals are combined following Mann (1975). 68 G. S. KMNTZ

200 200 roo roo

L P P P P

Figure 1.--Upper dental profiles of Figure 2.--Upper dental profile of crown areas of genus Homo. The upper Australopithecus africanus, Sts 52. curve is the erectus ?iEage with There is a greater relative emphasis neandertals substituted for the on the posterior dentition than in first inc.isor. The lower curve Homo but it shows a similar "leve1ins" is a pooled sample of sarriens. FTftu canine and premolars.

Figure 3.--Upper dental profile of Figure 4.--Lower dental profiles of Australopi-thecus robustus, OH 5 crown areas of genus Homo. The upper (ttZinjanthropus"). The posterior curve is the erectus ffi-rage with dentition is relatively large, and neandertals for the in- "ffiilTted unlike Homo there is a marked cisors. The lower curve is a pooled ir,"re""fE size from the canine sample of sapiens. through the second premolar. 69

AUSTMLOPITHECINE BODY SIZES

a./ ./,

,/' tl.\f\ .rt---- c\)l) t^ zoo / ,'"'----- / ..,"' / ..,"' i

/.r,'.t.,t' lo0 ,y',.'

',4

P

Figure 5.--Lower dental profiles of Figure 6.--Lower dental profiles of Austra.lopithecus af-ricanus, Sts 52, Au-stralopith.ecug robustus, Natron, solid solid 1ine, and MLD 18, dashed line. line; Sk 23rdashed line; and Sk 34rdotted There is a greater emphasj-s on the line. The posterior dentition is rela- posterior dentition than in Homo tively 1arge, and unlike Homo there is a but they show a similar "levffig" marked i-ncrease in size fiffi-the canlnes of the canine and premolars. through t.he second premolars. *t( J- The difference between the complete dental profiles of Aust.ralopith-ecus and Homo is in the anterior vs. posteri-or emphasis. In si-mplest terms this can Ue qGTFified by expressing the crown areas of the incisors as a percentage of the molar areas. For the upper dentition in genus Homo thi-s amounts to 33.77. in saplens and 35.0% for eregtus as shovm in Figure i. By more conventi-onal group- ing, neandertals are 39.47. and the usual erectus are 3L,47". A11 these upper dentition figures are closely grouped rttaTeiEE on about 35%. This is in marked contrast with the values for the two complete australopithecine upper dentitions of 14.0% for O H 5 (ttZinjanthropus") and 22.3% for Sts 52 (Figures 2 and 3). These are two extreme specimens which average about 1-8%in contrast to the 351l of genus Homo. In lower denffions the j-ncisor crowns are a consistently srnaller per- centage of molar cro\,rrrs because of the difference i-n t.he first incisors as noted above. For genus Homo in Figure 4 the lower j-ncisor percentages are 20.4% in sapl-ens, 2L.0% i-n,er.ectus as used here, 25.L"1 j-n neandertals, and L8.87. ln erectus proper. These lower dentition fi-gures center on about 2I%. The austrZl@ecine lower dentitions measure 19.77. for Sts 52, 12.77" for MLD 18, 11.8% f.or Sk 23, 10.0% f.or Sk 34, and B.B% for Natron (Figures 5 and 6). These average L2.62 ln contrast to the 277" of genus Homo. This incisor to molar ratio shows a marked .6iE7ast between the two genera but there are too few measurable specimens of Australopithecu_s to show a clear division into different forms within that genus. Two degrees of molar emphasis are at least suggested by the data, but it is only a single specimen that constitutes the least molarized type for both uppers and lowers. A greater number of usable specimens can be obtained by limiting the conErast to the second lncisor and first molar alone. This also permits upper and lower dentitions to be compared almost indiscrlminately. This comparison suffers from the fact that since only two teeth are involved, the results can 70 G. S. KRANTZ

oore easily be skewed by individual tooth anomalies, interproxirnal attrition, and measurement error. The upper second incisor of sapiens is 39.77" of tlne first molar, in erectus it ls 47.9%. In Australopi-thecus these average only 237"buE with great variarion (Sk 27-32.12, Srs 52-26.9%, Sk 55-20.5%, TyI L512- 19 .9%, OH 5-L9 .5%, Sk 52-18 .8"/"). The lower second incisor of sapiens is 32.27" of the first molar, ln erectus it is 35.7%. In Australopithecus these lowers average onLy 25%, again wGtr-considerab1evariati@,Sts24-33.0%,MLDI8-28.o7.'SkB45- 22.4%, Sk 23-20.62, Sk 34-19.9%, and Natron-l5 .6%). Now a distinction begins to emerge between two degrees of molar emphasis among the australopithecines but without the tight elustering of values that rnight be hoped for. This should be expected when only two teeth are being compared, with their possible individual peculiarities. As a test of these peculiarities I compared the sizes of all matching teeth from opposite sides of the same austral-opit.hecine jaws. In 93 cases the differences in crown areas ranged from nothing to 337", and averaged 4.67!. There is also the possibility that there are allometric differences in these incisor to molar ratios which depend on the absolute sl-zes of the specimens. Also the allometry rnight be different in the different forms. This cannot be escablished until the two forms are separated and the specimens properl-y assigned. The most consplcuous contrasts among the various dental profiles relate to the contours in the canine and premolar region, hereafter referred to as C-P-P. Two distinct types occur. In the 'rconti.nuous" type the entire dental profiLe forms a shal-l-ow "S" curve with no prominent jumps in size anywhere be- tween adJacent teeth. The C-P-P series increases evenly, being part of an almost straight line of gradual size increments from the second incisor to the second mol-ar. This "continuous" type seems to correlate lrith the hlgher degree of molar emphasis and is found in the skuIls usually called "robustr'. The contrasling dental profile might be called the'rleveled" type, where the C-P-P line on the graph tends to be roughly horizontal. This is the same condition as found in both species of Homo, where the canines and premolars are all nearly the same si-ze. Cornparedwith a "continuous" dentition of t.he same size, thettlevel-edtt one has a much larger canine and a much smaller second pre- molar, but otherwise its curve is often very much the same. It is as though this three-tooth segment of the dental profile has simply tipped with the first pre- molar forming the hinge and being relatively unaffected in size. The "leveled" dental profiles do tend to show less molar emphasis and have been associated with the so-called "gracile" skulls. Australopithecus dentitions that are sufficiently complete can be divided into these t\do types of "l-eveled" and "continuous". Figures 2 and 5 illustrate upper and lower dental profil-es of the "leveled" type which will now be referred to as africanus. Figures 3 and 6 illustrate the "continuous" or robustus dental profiles. Further specimens can be classified into these types on the basis of the following five tooEh size ratios.

Leveled C-P-P Continuous C-P-P

I,2toC Canj-ne about twice as large Canine half again larger C toPl About equal in size Premolar half again larger Pl to P2 About equal in size Second premolar half again larger P2 to Ml Molar about twice as large Molar half again larger C toP2 About equal in size Premolar about twice as large

(It should be noted that I am not using the usual paleontologistts pre- molar designations of P3 and P4 which are based on the assumption that these 7L

AUSTMLOPITHECINE BODY SIZES

correspond to the last two of the four premolars in the ancestral placental mammal dentition. I am tnclined to agree with Bolk (19f6), that our premolars acEually correspond to V2 and P3 of the original. To avoid the dilermna I simply nurnber them as they occur.) The canine Lo second premolar ratio should logically carry more weight than the others as i.t reflects both of the two differences in tooth si"ze be- ttleveledttandttcontinuoustt tween dentitions, while the other ratios each mea- sure only one of thern. In addition, the incisor to molar ratlo could be added as a sixth criterion. No one of these ratios by itself should lead to a posi- tive classification of a dentition as africanus or robustu_s. Any two ratios ought to be a fairly certain identification as long as no other data are eontra- dictory.

Individual Identif ications

By using the above criteria the various full and partial dentitions can be classified into the two types with a fair degree of confidence. The fol- lowing australopithecine specimens from South Africa can be identified as africanus on the basis of the ratios indicated after each.

L1ST A

TM l-511 (Sterkfontein) p1-p2, P2-ML TM 1512 (Sterkfontein) all ratios (uppers) MLD 2 Pt-Pz, P2-M1 MLD 11/23 all ratios (uppers) MLD 18 all ratios (lowers)

D-pp-M TM l-600 (Kromdraai) 'L '2' '2 "L Sts 7 c-Pr: PtlPz: Pz-Mt' c-Pz 1"r-J-1 Sts 17 P--P', P'-M- Sts 52 all ratios (uppers and lowers) sk 6/100 Pt-rz, P2-M1 p-p p- sk 12 -J, -2' -2',Mt sk 27 r--c, c-P- sk 68 on morphology (see below)

Likewise, Lhe following specimens can confidently be classed as robustus on the basis of the indicated ratios.

List A (continued)

TM 1517 (Krorndraai) ,t-rz, P2-Ml rt-D D o MLD 40 -D -Ml , C-P "r'Li 'L?'2i ' 2 2 sk 13/14 P*-P-, P--M* sk 23 all ratios (lowers) sk 34 all ratios (lowers) 72 G. S. KRANTZ

List A (continued)

sk 46 pL-pz, ,2-Ml sk 52 p1-p2, P2-M1, r2-Molars sk 55 tz-c, c-YL sk 63 C-P2t C-Molars sk 65 f -4, c-Pl, pL-p2, c-Pz sk 83 p1-p2, P2-M1

There are a few suprises in these assignments which contradict what is generally accepted. Sk 27 is the first such exception I noted several years ago and can nohr be made into a test case. If this one can be conclusively demon- strated to be an africanus from a "robust" site, then the rest should be easier to accept. Accordingly, a disproportionate amount of attentj-on will be paid to this specimen. Sk 27 is a crushed juvenile skull showing permanent upper inci-sors, can- ines, premolars, and first molars, only some of which are measurable. The crorin area of its canine is 86.7% of the area of its first premolar. This contrasts sharply with the corresponding measures of other Swartkrans indivj-duals: Sk 48-57.27.,' Sk 55-57.2"/", Sk 65-65.9%, and the peculiar Sk 83 with 77.67". It is comfortably among those from Sterkfontein which are TM L5I2-78.27", and Sts 52- 87.27.. This canine to first premolar contrast has long been recognized as probably the most diagnostic trait distinguishing the two forms (Coon, 19622273). Yet here a Swartkrans specimen is elear1-y of the africanus type. In more detailed tooth rnorphol-ogythe africanus affiliation of Sk 27 continues to hold true. This is evident in the following quotations from Robinson (1956). In describing the Sterkfontein upper canines he says (p. 44-45) z "The J-ingual grooves are not as c1-earl-y marked as in the case of P. r. crassidens and instead of converging sharply onto the gingival- emii6nce are almost paralleL. " This may be compared wlth the same in Sk 27 (p. 43) z "Neither of the lingual grooves is clearly defined but are merely small depresslons half-way down the crown and are widely separated (7 rnm)." Although the wording is different, the descriptlon of Sk 27 j-s essen- tially that of the Sterkfontein specimens. He says further about the upper canine (p. 45)z "The prominent swellings between the lingual grooves and the cervical line, whi-ch are so characteristic of P. r. crassidens, are not present in the Sterkfontein specimens." And of Sk 27 1p. 3): "The usual two swollen ridges near the cervical line are absent." Again Sk 27 is being described as a Sterkfontein specimen. Of the first upper premolars from Swartkrans (p. 55): "Usual1y buc- cal grooves are present in the form of two slight depressions runni-ng approxi- mately half-way from the occlusal surface margin toward the cervical line. These nray be equally developed, but more cofiunonly the mesial depressi-on is more strongly developed. In one case only is there a trace of an aetual groove present and this is found in the mesial buccal groove of Sk 27." Of these prernolars from Sterkfontein he says (p. 58): "The buccal face in all but two instances has well-defined brrccal grooves, the mesial one being more strongly developed than the distal one." 73 AUSTMLOPITHECINEBODY SIZES

200

P

Fi-gure 7.--Upper dental profiles of Sk Figure 8.--Lower dental profiles of 27, soLid line, and MLD LL/ 23, dashed Sk 6/100, sol-id line, and TM 1600 Llne. Both dentitlons show the (Kromdraai), dashed line. Both show africanus pattern, thus reclassi- the africanus pattern and are here fying Sk 27, reclassified.

200

ro0

c P P

Figure 9.--Lower dental profile of Sk Figure 10.--Lower dental profiles of 12 showing its afri-canus pattern in MLD 40, solid line, and Sts 51, dashed the premolars and first molar, and line, which show greater similarities Leading to its reclassification. to t,he robustus pattern in the canine through premolar slope * * So Sk 27 is like Sterkfontein, but one also wonders which were the other two exceptions alluded to from Sterkfontein. On the lingual face of the first upper premolar (p. 56): "This is not always the case and in Sk 27 the reverse is true." The second incisor of Sk 27 is also unusual for its type (p. 27). These descriptions are all rat.her convinsing, t.he measurements agree, 74

G. S. KRANTZ and there are no morphol-ogical contradictions. Unfortunately it simply did not occur to Robinson, or to anyone else at the time, that there could be any exceptions to the rule thaL each site contained just one type of australopith- ecine. Wallace (L973) also noted the peculiarities about. Sk 27 which set its teeth apart from other Swartkrans specimens without drawing the conclusion which now becomes clear when seen in a broader perspective. Morphological observations on specific individuals are rather unconmon, but a fewmore of them are illuminating. tr{e might note Robinsonts descrlption of the upper flrst incisor (1956:25) which states, "A distal lingual groove is present in each of the Sterkfontein teeth but the mesial one appears to be absent. " And of the Swartkrans teeth he states (1955 223) , "A11 the specimens have these two lingual depressions which are in the form of clearly rnarked grooves--the mesj.al one is missing in the case of Sk 68." It would follow f'robust" that Sk 68 rnay also be an africanus from a site. The "Meganthropus" j"r (or Sangiran 1941) is another that fa11s into this category. Because of its large size all attempts to fit it into Australopithecus have concentrated on trying to rnake it a robustus. Von Koenigswald (1973) put it sirnply, "In certain respects the lower jaw of Meganthropus combines charac- teristics of A. africanus (premolars) with those of A. robustus (size). " IIad he known otfrer aTffGuJ-sp-cimens were also quite farge-fG proper identifi- cation would have teen easy. The "Meganthropus" premolars are about equal in their cror^rn surfaces (118. L7 and LL6.62 *o2j ana the first molar is far larger (2OL.Z8 rw?). Thus i.t shows two good ratios of the "l-eveled" dental profile. The simple design of its premol-ars, especial-ly of the second one, is also an africanus trait. The sixth cusp, or tuberculum sextum, of the first molar has be6-Ififfi6d as a robustus trait (Robinson, L956), but its presence in the Taung specimen raises some uncertainty. In any case, the sixth cusp is more likely related to the absolute size of 'h4eganthropus" than to its phylogeny. (Sirnilarly, the high frequency of fifth cusps today on the Lower molars of Austral-ian Aborigines and some American Indians reflects their large teeth and not any especial-1-y close relationship. ) In order to show the australopithecine affinities of "Meganthropus" its detail-ed similarlties were shown to a jaw from Swartkrans, Sk 6/1-00, and one from Kromdraai, probably TM 1600, (Robinson, 1973). These comparisons are concl-usive indeed, but what they actually prove is that the two South African specimens are also africanus. Both of these "robust" jaws have already been shown to have the "leveled" type of dental profile in conLrast with most specimens from thej.r respective localities. The other "Meganthropus" jaws cannot be fitted into africanus. The specimen Marks found Ln L952 is too poorly preserved to classTfy, though it may well be of this type. The earlier specimen, Sangiran 1939, has already been shown to be more likely related to the oranguten (Weidenreich, 1945, and Krantz, 1973b). At this point we have seen that five South African speeimens which were previously classed as robustus because of their provenience actually turn out. to have africanus rorpE61ffi- This number may be compared with the ten 'trobust" specimens that can be equally positively retained in the robustus category on thebasis of their dental profiles. Arnongthose jrr" tr"dition"lly identified as africanus by site or size only one has so far been shifted into robustus with some assurance on the basis of its dentition. There remain eight South African "gracile" individuals which can positively be retained in africanus. So out of 24 identifiable specimens, 18 traditional assignm"rrt" and 6 "iJ-E6iilFred, are changed. 75

AUSTMLOPITHECINE BODY SIZES

On a somewhat lower level of certainty, a few additional jaws can be assigned to the tr^ro dentaL types as probables. The criteria are the same tooth size ratios as used before but these judgments are made on the basis of just one clear ratio or on two that are aL least strongly suggestive of one type as agai-nst the other. A few of these assignments night be incorrect.

List B

Probable africanus ProbabJre robustus

MLD 6 pL-p2 Sts 51 C-P,, MLD 9 P2-# sk 11 p2-ioltt" Sts 1 P1-molars SK 48 c-P1 Sts L2 pL-p2 sk 49 p2-Ml Sts 42 pL-p2 sk 845 p1-p2

This list introduces only one switch, the "gracile" Sts 51 being classed with robustus. Combining lists A and B we have some new totals. There are 18 I'robustr! identifiable africanus specimens, 13 from "graci1e" sites and 5 from ones. There are 16 robuslus specimens, 14 from "robust'r sites and 2 from "gracile" ones. Seven out of 34 specimens, or over 20%, are here considered to have been misclassified in the past because provenience was given prlority over morphology j-n their assignments. If the remaining dental- specimens, which are not easily classifiable here, show anywhere near these same proportions, then it is cl-ear that site allocation ls highly inaccurate. Up to now aLl pubLished comparisons of the tvro forms of australopith- ecines have been made between samples pooled according to their provenJ-ence alone. This ls rather like comparing measurements of two samples of dogs and cats, one group which l-s labeled "dogs" and contaLns L47"cats, one group which is label-ed "cats" and contains 28% dogs. It could be concluded there is no difference between "dogs" and "caLs" in view of the variability found in each sample! In the case of Australoprghe"u", all future comparisons should be restrictedtothosespe.ffiidentifiedindividua11yontheirown merits. Even if Sk 27 were the only wrongly classified specimen that we knew of for sure, the possibility that there could be more should be enough to invalidate any pooling of specimens by site. The assignments to species made here are based mainly on dental profiles and in only a few i-nstances could other rnorphological traits be used. A dir- ect sorting of the specimens themselves might change a few of these assignments, and this should be done. Preferably such an original sorting should be made by a marmnalian paleontologist who is not familiar with this material and could nor recognLze any of the individual specimens or catalog numbers.

Sexual Dimorphism

Sk 27 is notable for more than being in the "wrong" site; it is also one of the largest specimens known. No doubt one of the reasons Sk 27 was not recognized earlier as an africanus was because it was so large, as well as being found in a "robgst" site. The combined crown area of j-ts canine and first premolar Ls 237 mrn2, while the fou4 other Swartkrans specimens roith both of these teeth present range from 196 r*2 to 2L0 mnl. Even Wolpoff's (1971) "robust" sample, includlng large east African specimens, gives an average of only 225 mmz for these two upper teeth. 76 G. S. KRANTZ

The .great size of Sk 27 suggests it is a male while most other africanus specimens are females. This possibility becomes more clear with the recognition of other africanus as males. These are Sk 6/L00, Sk 12, TM 1600, and MLD 2. Each of these is singled out on the basis of its larger size in contrast to the majority of afri-canus specimens. The "Meganthropus" jaw now makes complete sense because its size, which had caused many to try to class it with robustus, j-s actually quite normal for a male africanus. Some inkling of this size discrepancy within africanus and also the scarcity of male specimens was indicated by Robinson (1956;ABt in his thorough description of the australopithecine teeth. One of the canines from Sterk- fontein (Sts 3) wasrfar larger than the nine others, uBper and 1ower. Its croriin area is 127 mnz while the others range frorn 75 ,*z to L02 mn?. If Robin- son is correct that Sts 3 is a loroer canine, its upper equivalent would be even larger. He drew the obvi.ous conclusion that this is the only male specimen out of a sample of 10 individuals, the other nine being females. Robinson (1956:93) also noted the large size of the upper third molar in Sts 28/37. This indicated a single male as against six very much smaller female individuals for this tooth. In all other Sterkfontein teeth either the di- morphism did not show or else there were no more males. These indications of a great sexual dimorphism and shortage of males were not followed up by Robin- son or anyone else. The picture is now clear that there are two very different sizes of africanus with the much larger specimens being considerably less common than the others. Unless one postulates two species with this same dental rnorphology Ehey must represent two sexes. The previously identified africanus individuals from South Africa may be divided by sex as follows.

List C

Male africanus Female africanus sk 6/100 TM 1511 Sts 1 sk 12 TM 1512 Sts sk 27 MLD 6 Sts L2 MLD2 MLD 9 Sts T7 TM 1600 t"LD LL/23 Sts 42 MLD 18 Sts 52

It would be tempt,ing to include the big canine of Sts 3 in the male column and the other nine Sts canines with the females. However, at least one of Ehese nine, Sts 51, is a _Eglus_!-gE-and it is not determinable how many more may be as well-. Likewise the upper third molars of Sts 28/37 is a ma1e, and the other six include at least one robustus, TM 1517, and maybe more. The unusual sex ratio of 5 males to 12 females would be exaggerated if these other Sterkfontein dentitions were included. Ruling out only the already known robustus specimens, one possible robustus (Sts -54), and those already counted, and by combining TM 1512 with TM 1561, we get eight more specimens as follows. List D

Male africanus Female africanus Sts 3 Tt't 7527 Sts 48 77

AUSTMLOPITHECINE BODY SIZES

List D (continued)

Sts 28137 Sts 2 Sts 50 Sts 36 Sts 53

Combining lists C and D gives a sex ratio of 7 to 18. This may be too high as more robustus rnight be included in the last group of six females that were just added. In any case the sex ratio is abnormal and will be dealt with 1ater. If a rnajor sexual dimorphism exists in africanue it night be reasonable at least to look for a similar size discrepancy in robusEus. It is there and has been known for a long time but has not been recognized for what it is. Again, there is a key specimen that serves to make the point and the rest is easier. This is Sk 15 which was originally known as "TelanEhropus" and more recently has been put in genus Homo by most workers. Broom and Robinson (1949) considered the possibility that Sk 15 rnight be a female robustus egtanfttrggge. crassidens in Broomrs earlier terminology) but rejected this notion rnainly on the grounds of its sma1l size. Those who think it is a female robustus include Dart (1955), Wolpoff (1968, 1970) and others. Those who now Lreat it as early Homo include Howell (1969), Clarke et al. (l-970, 1-972), Robinson (L972), and many others. Part of the problem here relates to t.he second "Telanthropus" jaw, Sk 45. Other than for its srnall size there is nothing to associate it urorpho- logically with the original "Telanthropus" jaw, Sk 15. Their contrasting fea- tures can easily be sunnnarized.

sk 15 sk 45

Body of marrdible thick relative Body of mandible thin relatj-ve to its height. to its height. Molar roots wide relative Lo Molar roots narrow relative to

CfO\^7nS. CrOIiInS. Third molar similar in size to Third molar much smaller than other molars. other molars. Coronoid covers third molar. Coronoid behind third molar. Lower margin horizontal. Lower margin dips anteriorly. Coronoid far lateral to molars. Coronoid close to molars.

A11 the listed traits of Sk 15 are australopithecine in general and can be confj-rmed from published illustrations (Robinson, 1953) and the cast. Those of Sk 45 aLI point rather to a more human morphology, presumably Homo erectus from the lack of any indication of a chin. The error has been in assuming these two jaws could be used together in describing their type. When the original specimen, Sk 15, is considered on its own merits it is basically australopithecine in spite of its small size. A11 the teeth in front of the molars are missing, but their sockets are preserved well enough to make rough estimates of their sizes. The spaces available for incisors and canines are rather small while the premolar sockets are clearly too big for erectus sized teeth. Even an approximate reconstruction gives an australo- 7B G. S. KMNTZ

pithecine contrast between incisors and molars, and the gradation of si-zes in the C-P-P series is definitely robustus. One loose tooth, a first premolar, is not diagnostic as it could fit an africanus reconstruction just as well, but the canine and second premolar sockets would not permit this. The jaw is clearly that of a remarkably srnall robuslus in all essential traits. Its short ascending ranus will be d"alt-iith--1fter. If Sk 15 is a typical female robustus then there should be more such specimens, and there should not be any clear sexual di-morphism among the lar- ger individuals. Robl,nson (1953:484) thought he san some indication of sexual dirnorphism among the regular Swartkrans sample but adruittedly it was slight and poorly rnarked. His data do not prove anything more than chance distribu- tion of sizes within a single type. The most obvious additional female robustus would be the maxilla, Sk 80, comruonly grouped with the other specimens as anottter "Telanthropus". This maxilla has now been matched with a partial skull, Sk 846/847 (Howell, 1969), which had long been classed as an obvious robustus of unusually srnall size (Robinson, Lg6O and. L967; Tobias , 1967). Si"ce Sk 80 has been fitted together with Sk B46/847, many have reclassified the enti.re specimen as an early example of genus Horno (Clarke and Howell, L972), while others dissent and find this sufficient proof that the whole lot belongs in robustus (Wolpoff, 1970). I must agree with WoJ-poff that if this specimen j-s separated as Homo, then all previous descriptions of robustus skul1 morphology become meaningless. Another robustus skull from Swartkrans, Sk 48, is also very smal1. Pilbeam and CoutE-lT97Z] describe it as being even smal-ler than Ms. Ples, the best known Sterkfontein skul1, and simply conclude that, "Comparisons should be between samples." Its morphology is robustus, its teeth fit the "continuous" curve and are very small, so its designation as female robustus would appear to be natural. Of the various dentitions assigned earlier in this paper to robustus a few rnore can be selected on the basis of their sma1l teeth as being female. These are Sk 11, Sk 46, and Sk 65. A11 identifiable robustus dentitions may be listed according to their probable sex.

List E

Male robustus Female robustus Sts 51 sk 49 sk 11 TU 1517 sk 52 sk 15 MLD 40 sk 55 sk 46 sk 13/14 sk 63 SK 48 sk 23 sk 83 sk 65 sk 34 sk 845

The sex ratio of these identified dentitions is then 12 males to 5 fe- males. This is skewed in the opposite direction from that found for africanus and to a similar degree. Sk 80/846/847 was not added at this poi-nt because it was identified by means other than its dentition, and similar procedures will be used to categorize many other specimens shortly. Outside of South Africa some specimens of robustus morphology can be identified and sexed, which in turn will clarify one remaining problem with ttTelanthropus." "Zinjanthropus,ttor OH 5, is a clear robustus male on the basis 79

AUSTMLOPITHECINEBODY SIZES of its dentltion. Two recently discovered and undescribed skulls from Kenya are clearly robustus frorn their casts, ER 406 being a male, and ER 732 being a female (see also Tobias, L973, and Holloway, L973). The male skulls are somewhat Larger than tl-reir South African counterparts, so the same rnight be expected for the females. The female ER 732 is about Lhe same sj-ze as Sts 5 (Ms. Ples) , being larger ln some measurements and smaller in others. Accord- ingly, the South African female robustus would be expected to be smaller than either of these skulls. The sr"[GTi6-of "Telanthropus" fiLs these expecta- tions. The remaining problem with "Telanthropus" is the short ascending ramus, reconstructable on Sk l-5, which contrasts strongly with the other tall jaws from that site. The short ramus was interpreted by Robinson (1953) and by Clarke and Howell (L972) as indicating a braincase larger than those of the other Swartkrans skulls. Their reasoning was that the short ramus meant the base of the braincase was lower, and hence presumably it also extended upward and in other directions to a similar degree. It apparently did not occur to then it could just as easily indicate the jaw was simply higher. The more bulbous shape of forehead of Sk 80/846 /847 was also taken to indicate a larger braincase and higher evolutionary status (Clarke and Howell, L972). These conclusions are unwarranted because within a species, a shorter ramus does not correlate with a larger braincase but rather with a smaller one. One need only compare female gorillas with males to see Ehis. In the females the ascending ramus is absolutely and relatively shorter than it is in the males, yet the braincase is significantly srnaller, not larger. Of course if "Telan- thropus" were in fact another species, then its braincase rnight be of any size, but the r€rmushetght does not prove this any more than do its srnall teeth. The forehead ls likewise expected to show more height in the smaller brained, but even smaller faced, females. The argument that "Telanthropus" rnorphology shows them to be female robustus is internalLy consistent and admittedly circular. But the argument Eh"t they are larger brained Homoerectus is equally circular. The solution to this comes from the fenale robustus skull, ER 732, from East Africa which is complete enough to include the upper dental arch and the glenoid fossa so the ascending ramus height of its missing rnandible can be measured. This height l-s very little and about matches that of the "Telanthropus'r mandible Sk 15-- they would almost fit. In an even smaller South African female robustus skull there would be no problem. There remain no morphological features to exclude the 'rTelanthropus'r specimens from being norrnal female robustus. The Sk 45 jaw fragment now stands alone as the only specimen frot--SffiE[rans that can be attributed to genus Homo. One more female robustus should be mentioned here, the childts jaw from Kromdraai, TM 1536. Inlallace (L973:21) noted it has the most molarized first lower deciduous molars, but its first permanent molar is one of the smallest in the site. This is a contradiction only if one accepts the pre- mise that large size and molarization go together.

Cranial Distinctions

Up to now this discussion has centered rnainly on the dental profiles of the South African specimens to distinguish the two species. Other material, by anatomy and geography, has been used to help clarify this core sample but has not yet been counted in any of the totals. By contrasti-ng the skulls with classifiable dentitions some cranial characteristics of each species can be determined. Those crania and jaws with afrieanus dentitions include Sts 1, 80 G. S. KRANTZ

Sts 7, Sts 17, Sts 42, Sts 52, MLD 2, TI4 1511, and ER 732. Sts 5 can be added to this list on the basis of tooth socket sizes and spacing. The cranial material with robustus dentitions include Sk 23, Sk 34, Sk 46, Sk 48, Sk 49, Sk 52, Sk 63, TM 1517, ER 406 and OH 5. A comparison of only these skulls and parts provides us wiLh a list of structural contrasts not greatly different from those generally reported. The excepti-on is that absolute size, and allometric changes consequent on size, are not being considered here. For example, the sagittal crest is dependent on Lhe size and especially the length of the temporal muscles which relate directly to body size, and on their at- tachment areas on the brai-ncase which varies only slightly with body size. Most of the distinctions between the two skull types can be described under the general heading. of the cranial base being opened-out as in africanus or closed-up as in robustus. In comparative terms the africanus- typ"-shEils-EE-e following characteristics.

1. The nuchal crest is relatively higher, and is also higher along the sagittal arc of the occipital than it is in robustus.

2. The foramen magnum is relatively farther back on the base of the braincase.

3. The basioccipital length (basion to hormi-on) is greater.

4. There is more lower facial prognathism.

5. The face "hinges" in the midorbital region, the forward inclination of the lor^rer part making a conspicuous angle along the lateral margins of the orbits.

6. Consequently, the entire underside of the sku1l from inion to prosthion is a rel-atlvely great distance.

Additional traits of africanus vs. robustus not obviously related to the cranial base include:

7. A more sloping mandibular symphasis.

8. A more prominent nasal area.

9. Less anterior projection of the malars, and a less advanced attach- ment of the masseters onto the maxi11a.

10. A somewhat lower hafting of the face onto the front of the braincase.

11. A seemingly more elongated braincase, especially as viewed from above.

It is interesting that in the first six characteristics robustus is more like rnodern man in having the more closed-up cranial base. A11 structures from prosthion to inion are packed more closely together and shortened. This does not indicate any especially close relationship between robustus and ourselves. A similar kind of contrast occurs between chimpanzee and orangutan skul1s. A1- though both apes are more opened-out in their cranial bases than either of the australopithecines, the orang is more so. No special relationships are implied here either, but a similar eause rnight be looked for. B1

AUSTRALOPITHECINEBODY SIZES

The more nearly vertical chin region of robustus is also more like that of recent man, but there is no connection" This follows from the smaller in- cisors of robustus and its relatively more massive lower mandibular body. The relnai.ling characteristi-cl, well as dental profiles and tooth "" morphology, all relate africanus more closely to genus Homo. One may wish to postulate both forms as human ancestors with a subsequent breakdown of the presumed reproductive barrier. As this seems unlikely on present evidence, a choice must be made between the Lwo forms and I would think africanus is far more likely to be our ancestor. There are other, more detailed, differences buL the above list ernpha- sizes those which are most easily seen in the published illustrations and i-n casts, and t.hey are €rmongthose most frequently mentioned in the litera- ture. These characteristics serve to classify some additional skull material that is either without teeth or without sufficient published description of what is there. These cranial distinctions assist in classifying a number of specimens including some East African individuals which have not yet been discussed. The most signi-ficant of these is ER 1470 announced by R. Leakey (1973) and partially described by Day et a1. (1975) as being an early representative of genus Homo from East Africa, Despite its large endocranial capacity and somewhat human-like appearance Wells (1973) correctly diagnosed it as a large australopithecine similar to africanus in design. Much of its pedomorphlc appearance may follow from the possibility that it is a juverrile with a den- tal age, in modern terms, of about L2 years (Krantz, L974). The spacing of its tooth sockets appears to fit the africanus pattern, but this could be that of genus Homo just as easily. The relief on the extant part of the palate and the location of the transverse palatine suture (Day, et al., L9752464) both indicate the palate ended with just enough room for the second molar and no more. If this is correct, and if the individual had grovrn to maturity, then it would have been considerably larger than OH 5 ("ZLnianthropus"). In its foramen magnum posit,ion, facial prognathisrn (in the first, correct reconstruc- tion), nasal height, and angled lateral orbit margins, it conforms perfectly to the africanus design. Since extremely large male africanus specimens have already been identified on other grounds, the identity of ER 1470 seems clear. The endocranial capacity of ER 1470 has been recently measured by Holloway at 775-780 cc. and this would have grown to over 800 cc. with mat- urity if rny age estimate is correct. This size is far beyond the 530 cc. of OH 5 which is the highest figure generally accepted for any australopithecine. Perhaps this will not seem too far out of line when it is remembered that most africanus remains dealt with so far are females, and the dimorphisn seen in the jaws and teeth is remarkably great. If Sts 5 is a typical South African female africanus with 485 cc., a gorilla-like sexual dimorphism would give about 570 cc. for the males (females being 85% of males according to Tobias, L973). Another L5% difference between South and East African forms (if true) would bring the expected size up to about 670 cc. From this to just over 800 cc. is within the range of individual variation. This is quite a series of jumps in size, and its beginning step of 485 was already the largest of the fernale africanus. A reconsideration of the "habilis" skulls points to the solution of the endocraniaL size problem. Three of these have been carefully reconstructed and corrected to adult values of OH 7--684 cc., 0H L3'-652 cc., and OId 16-- 633 cc. (Tobias, L97L). Their average of 657 cc. is remarkably close to the 670 cc. calculated above as the expectable size forEastAfrican male africanus. B2 G. S. KRANTZ

At least one of these, OH 7, also has a male africanus type of dentition. To regard these all as normal male africanus for their area requires no great imagination. ER 1470 may be classed with them as an unusually large or en- cephalized individual. The large male africanus should not have been difficult to identify but for the fact that all braincase material available so far is from im- mature individuals. They have the more pedomorphic desi-gn of youth, and this, combined with their large interiors, has naturally led to their being taken as more human than the other australopithecines. The notoriety that accompanies the discovery of early "man" can only contribute to this bias. One of the "habilis" skulls, OH 24 wLth 590 cc., was left ouL of this tabulation. Its teeth are small and from the cast could as easily be of the robustus pattern. Its cranial base is also very compressed from front to back, and otherwise has a female robustus morphology. The braincase recon- struction is by no means sat.isfactory and it should be re-examined to see if, as I suspect from the cast, it could be scaled down considerably. From South Africa the braincase of Sts 60, at 428 cc., has been com- bined with TM 1511 (Mann, 1975) which is a female africanus. Sts 71 with the same capacity is not immediately identifiable, and could just as easily be a female robustus, though I will here add it to the africanus 1ist, but with reservations. Sts 19/58 has a capacity of 550-570 cc. (Tobias, 1973> which would probably make it a male africanus as these capacities are being inter- preted hdre, but a male robustus cannot positively be ruled out. MLD 37/38 at 435 cc. (Tobias, 1973) is regularly taken to be an africanus because of its site, but as is now coming to be realized, this is not necessarily true. I will class it as africanus though it could be a female of either species. Sk 1585 is probably a male robustus with a capacity of 530 cc. (Holloway, L973). MLD L is only a parieto-occipital portion of skull but its size indi- cates a large capacity. From its published dimensions (Tobias, 1973) as compared with those of other australopithecines its capacity ought to be at least 600 cc. It is almost certainly a male africanus. A new tabulation should be made at this point of all the species and sex assignments just made, now incl-uding those frorn outside South Africa"

List F

africanus robustus Male Female Male FemaIe Sts 19158 Taung SK 1585 sk 15 MLD 1 Sts 5 OH 5 SK BO OH7 Sts 71 ER 406 TM 1536 oH 13 l,rLD37138 OH ')L oH 16 ER 732 ER 1470 ttMeganthropustt

The grand totals now from lists C, D, E, and F are 36 arrrcanus vs. 25 identifiable robustus. The africanus divide into 14 males ar,d 22 females, and the robust,ts-?TilEE-into t5 rnates ana 10 females. So the sex ratios continue B3 AUSTRALOPITHECINEBODY SIZES to be skewed, wi"th a shortage of male africanus and of female robustus. If the reader has so far followed ttre case for establistting ttrese four types of Australopithecus, two species with two sexes of each, then they will have noticed the relative sizes involved. In their dentitions each species shows a great difference in size between the sexes, but there is not a notable difference between the species. Past studies have shovm larger dentitions in robustus but this is partly because some specimens were wrongly assigned, and F;tit-6ecause of a very dif ferent sex ratio in the recovered specimens, For the most part we have been comparing male robustus with female africanus, and ignoring the other two categories. For endocranial capacities we have been doing much the same Lhing, com- paring male robustus with female africanus. The sex and species identifications developed here result in four groups of brain sizes as follows. These are given in South African terms with the East African data reduced by L5% before being pooled. Male africanus 572 cc. (6 speclmens) t"tat" rotG[is 47L cc. (3 specimens) Female africanus 443 cc. (5 specimens) Female rob"stus 430 cc. (1 specimen)

The sample sizes here are too small and some reconstructions Loo uncer- tain for these values to be taken very seriously. Also some of the identifi- cations may be incorrect and the percentage reduction of the East African capacities nay not be accurate. Nevertheless the size sequence is impressive-- males are bigger than females and africanus are bigger than robustus. Unless one postulates a different degree of enc.phalization for the tro it "p".ies can be suggested that body sizes are similarly distributed.

Body Size Reconstructions

The next step is to re-examine the postcranial bones on which body size estinates harre been made. There are probably no South African postcranials that can positively be associated with cranial or dental individuals. Close proximity in a deposit of scattered debris is no evidence for a single indivi- dual unless mosL of the body is present and parts of other individuals are not. The usual procedure in the past has been to put all postcranials from Swartkrans and Kromdraai into the robustus category, and all those from Sterk- fontein and Makapansgat into africanus. Generally these are large bones in the former group and srnaller bones in the latter. Where particular bone sizes do not fit the pattern, these are either ignored or dismissed as belonging to some other species. One consequence of this procedure was the tooth to body size rat.io Wolpoff (1973) rrorked out for africanus which sirnply didnrt make sense. He used the accepted species designations for the dental and postcranial spec- imens and found "gracile" australopithecines had 21.2 mn( of cheek tooth grinding surface for each kilogram of body weight. This r,uas twice the rela- tive chewing surface as in the chimpanzee and almost three times that of modern man. This was not too unexpected, but the equi-valent ratio for robustus, which he did not present, was surprising because it would have been substantially lower. By conventional allocation of specimens, robustus cheek teeth are a little bit larger than those of africanus, but ttreir l"aies are perhaps twice as 1arge. Yet individual robustus dentitions show a much higher emphasis on malarization than do africai.rE--in--their tooth rnorphology and relative cro\^7n 84

G. S. KRANTZ

sizes. Individual dentitions contradict the tooth to body ratios. The only way out, of this dilernrna is that the postcranial assignments are almost entirely w'rong. Since it has been shown here that the dentitions of the two speci-es are not site specifie, but are mixed in all four of them, then some other procedure must be devised to sort out the postcranials by species and by sex. This proved to be fairly si.mple when it was considered that four well separated body sizes hlere apparently involved. If the australopitheci-ne postcranials sorted themselves into four size cl-usters these could be directly equated with the four sizes indieated by the skulls. Before doing this it shoul-d first be establlshed that there are no locomotor differences between the two species that can be used to distinguish them. The only differences are in t,otal size and allometric adjustments to size. This has been convincingly shown by Lovejoy (L973 and elsewhere) and he ls supported in this conclusion by Wal-ker (1973) and McHenry (1975). The other vlew, that robustus was a less proficient biped (Robinson, L972 and Napier, Lg64), is*ElsEl-on these size distinctions and the tentatively recon- structed femur length of one individual. Because the larger types increase in weight (cube of linear dimension) faster than in strength (square of linear dimension) some leverage changes are required, and are found, which accomrnodate for this. The relatively greater ischial projection ln the larger pelvis is a classic example of this. Measurenent ratios are meaningless without consider- ation of -absolute size. A relat.ed example of this accounts for differences in femur head sizes and neck lengths. In most austral-opithecines these heads are smal1 and necks are long when compared with modern human specimens. Lovejoy and Heiple (1970) showed how this fol-l-owed from the same locomotor design as in man being com- bined with a narrower pelvic lnLet. Infant brain size is the rnajor determining factor in the size of the human pelvic aperture, so the srnaller brained australo- pithecines could afford the luxury of more closely placed acetabular sockets. This in turn permitted a longer lever of the femur neck putting less pressure on the femur head and thus al,Lowing it to be smaller. This is an altogether nore efficient arrangement thdn the one we have been forced into because of our big bralns. But the al-lornetrics of brain size would argue that a small australo- pithecine would not have it so easy. A very small female would have to be designed for births that are not correspondingly as small as herself. The ratio of pelvi-c size to infant head for females of either species might have been little better than that encountered today. Thus smaller femurs and pelves should be expected to be more like ours, while the larger ones would have the better locomotor design. My own reconstructions of body sizes here are based on published mea- surements of the postcrani.al bones wherever possible and on measurements of available cast,s. These have all been compared with a standard Hsrne sapiens skeleton standing 152 crn. tall and which should have weighed 45 t g. itt fif* The method used in these calculations allows for the fact that body mass increases with the cube of a linear dimension, while surface area normally i-ncreases only with the square. Length measurements of bones of variously sized individuals are not in direct proportion to their weights, but the cubes of these lengths are in proporti-on to body weight. Thus if one individual has a femur 102 longer than another, his body weight should be about 33% greater (the cube of 1.1 being 1.331). If body proportions remain roughly const.ant, this rule should apply to the lengths of all long bones, and to all partial measurements along the long axes of these bones as wel1. If the 85 AUSTMLOPITHECINE BODY SIZES lengthwise measure of part of a bone Ls LO7. less than the same measurement on the standard specimen, then the body mass should be only about 73% of that standard (the cube of 0.9 being 0.729). Any linear dimensions can be handled in this same way as long as the same parts are measured for each comparison. Thus the lengthwise measurenent of any australopithecine bone, oubed, rnay be compared with the s€rme measurement, also cubed, of a standard skeleton to get a ratio of their body weights. The procedure for transverse measurements is different. The weight bearing capacity of a bone shaft or joint surface should be in proportion to j-ts cross-sectional area, perpendicular to the line of weight transmission. These areas can be expressed as the product of two transverse measurements such as long bone diameters. In most cases simply the square of one such transverse diameter will accurately show its relative weight bearing capa- city. If the breadth of a distal femur Ls L0% greater than the standard specimen, iL shows a 2L% greater weight bearing capacity (the square of 1.1 being L.zL). Likewise, if a diameter is 10% smaller this means 19% less weight (the square of 0.9 being 0.81). This different treatment of lengthwise vs. transverse measurements reflects the redesigning of bones belonging to different individuals of simi- lar body builds but with contrasting absoluLe si-zes. In a series of bones laid out in order of increasing sj,zes, the cubes of their lengths will in- crease in direct proportion with the squares of their diameters. Applying this to the australopithecine fragments meant taking their measurements and those of the standard s_apiens skeleton, then squaring or cubing each pair as appropriate, and geLLing a ratio of body weights. For every fragment a number of measurements were obtained, a weight ratio calcu- lated for each, and these ratios \^rere averaged to settle on the probable actual body weight. The various ratios found with each fragment generally clustered together closely, but there were some notable exceptions. Ratios based on femur heads were too small and those based on femur neck lengths were too high to be correct, as would be expected from the locomot,or adaptations discussed above. These, and a few other erratic ratios from innominates had to be left out of the calculations. A regular discrepancy was noted between the results gotten from shaft diameters and articular surfaces of the same bones. Midshaft diameters gave about twice the reconstructed weight as the more realistic articular surfaces, and shafts near their ends gave about 20% xoo much. A11 reconstructed weights based on shaft diameters were reduced accordingly. Length measurements on proximal radii were obvi.ously erratic and it was decided to omit them alto- gether. There are still a number of possible sources of error in this procedure. The casts may not all be exact reproductions. The standard sapiens skeleton I used rnight be not quite of normal development. The various processes on the australopithecine bones are only assumed to be in the same relative positions. ' Normalization of a long bone (cube of length and square of diameter) doesn't hold perfectly for exLreme sizes. Australopithecine body proportions may differ from ours, though I doubt this. But these same problems, or similar ones, beset anyonets attempt to reconstruct body sizes. r+& Table 1. Summary of calculations used to determine body weights for 18 indivi- duals. Ratios show the linear measurements of the australopithecines in rela- tion to a standard sapiens skeleton. Numbers in the "Square or Cube" column are multiplied by 100 so they also read pounds of body weight. B6 G. S. KRANTZ

TABLE 1 Specimen Bone Ratios Square or Cube Kilograms

MLD 32 Prox. Rad. L.25 156 70.9

Sts 68 Prox. Rad. T.2L L46 ) 70.5 1.40(end shaft) .83 x ]96=L63 | ER 737 Femur 1.36(end shaft) .83 x 185=154 I 65.3 1.63 (midshaft) .5 x 266=133 J ? sk 82 Prox. Fem. 1. 19(length) - L42 64.5 sk 97 Prox. Fem. 1.08 rr7 ) 58.9 1.31(end shaft) .83 x t7I=L42 J

Sts 7 Prox.Hum. 1.11 L23 55.9

ER 999 Femur 1.10 T2T I 55.9 1.58(midshaft) .5 x 250=125I (Kr) TM1517 Dist.Hum. 1.03 106 ) L.27(end shaft) .83 x 161=134| 54.L Prox.ULna 1.08 LL7 ) sk 18 Prox. Rad. 1.11 r23 I 53.9 1-.17 (end shaft) .83 x L37=l-L4I

MLD ].5 Prox. Rad. l-.05 110 50.0

ER 471 Tibia L. 2B(mldshaft) .5 x 159=80 37.5 sk 34 Dist.Fem. .89 79 I 36.8 1.00(end shaft) .83 x 100=83 I

ER 803 Fem,Tlb, Ul. 1.26(rnidshaf ts) .5 x l-59=80 36.4 TM1513(Sr) Dist.Fem. .89 7eI 34.5 .94(end shaft) .83 x 88=73 ) 0H B(a) Talus .75 r s6 l Mrtarsals .83(lengths) 57' 25.7

Sts 14 Lumbar .73 ^ s3) . 82(lengths) r ss ( 2I.5 Thoracic '64 1 4L{ . 76(lengths) - 44)

0H6 Tibia .67 45 I 20.7 .96(rnidshaf t) .5 x 92=46 J

TI[517(Kr) Talus .62 3B L7.3 * B7

AUSTRALOPITHECINEBODY SIZES

These calculations are shotm in Table 1 which includes everything except the raw measurements. In two cases it was necessary to separate a specimen. The talus associated with the elbow parts of Tt"t 1517 is far too small to belong to the same individualn and the oH 8 cl-avicle is too rarge for the associated foot. The resul-ts of these calculations are graphed in FIg- ure 11. *

Speclmen Kilograms Species MLD32 'rrl Sts 68 ER 737 F* sk 82 62

sk 97 59

Sts 7 56

ER 999 53

TM 1517 50 africanus

sk 18 7

MLD 15 4

4I

ER471 3B ,u robustus sk 34 35 ]-

ER BO3 32

TM 151-3 29

oH8 26

Sts 14 23

oH6 20

TM 1517 L7

Figure 11.--Reconstructed body weights shornrnon a scale. The four groups are each averaged and labeled by sex and species as interpreted here. *** Four fairly well defined size clusterings are evident which center on 2l' 36, 55, and 68 kg. Each of these contains four or more specimens and no group is exclusively of specimens from either "gracile" or "robust" sites. These 88 G. S. KMNTZ can be equated one for one with the four cranial sizes that have already been established. The body weights of africanus would then be 68 kg. for males and 36 kg. for females, wj-th an average of 52 kg, and a sexual dimorphism of nearly 2 to L. In robustus the weighLs are 55 kg. for males and 21 kg. for females giving an average of 38 kg. and a sexual dimorphism of at least 2-L/2 to 1. These dimorphisms are comparable to that reported for baboons whose eco- logical circumstances have long been compared with the australopithecines. How- ever, the apparent reason in the case of the baboons, that of maximumfighting size in the males coupled with the most economical size for females, seems i-nappro- priate here from the lack of projecting canines. This degree of dimorphism might be expected in a descendant of nar"gth"S"". if this i-ndeed is merely the femal-e of Dryopithecus indicus, as has beet-suggested (Krantz, L973a). A simpler, but less exact, method of categorizLng these fragments against the st,andard skeleton would be to compare each of them visually and by feel with the corresponding parts of the standard skeleton. They can then be judged as being (1) rnuch larger, (2) larger, (3) smaller, or (4) much smaller, than the standard. This procedure was used for some specimens, especially j-nnominates, whlch invol-ved too many problerns for direct measurements. These subjective categories can be equated directly with the four weight groups and a number of specimens thereby added to them. These additions are:

List G

TM 1605 (Krorn.) Ilium Large Male robustus MLD 7 Ilium Very snall Female robustus MLD 8 Ischium Very smal1 Female robustus sk 50 Innominate Large I4a1e robustus sk 84, 85 Metacarpals Very small Female robustus

OH 8(b) Clavlcle Very large ldale africanus

This brings the total number of body size individuals to 24, of which 9 are here considered africanus and 15 as robustus. The total sex ratio is an almost normal 13 males to-IT*fema1es. t" gqri*gg there are 5 males and 4 females, and in robustus there are 8 males and 7 females. South Africa is the source of 17 of these postcranial individuals, 8 belng ttgraciletr ttrobusttt from sites and 9 from ones, and the remaining 7 are from East Africa. Each of the size categories includes one or more individuals from each of these three sources. Of the 17 South African specimens 10 are from either "graclle" or "robust" sites as previous classifications would have named them, and 7 specimens are changed from Lheir traditional categories. The most famous of these is the tiny partial skeleton of Sts 14 which has always been taken as the classic of the "gracile" postcranials. Just like the small- est of the known skulls, this Sts 14 is a female robustus. Other attempts to estimate australopithecine body sizes may be more pre- cise in some instances but these generally have not applied a consistent meth- od so directly to the data. McHenry (L974) gives the most comprehensive recent set of reconstructed body sizes, most of which compare well with those given here after his statures are converted to weights. If body weight is taken as the goal, my method goes directly from bone measurements to weights in a single step. McHenryts method goes from measurements to reconstructed bone lengths, from there to stature reconstructions, and from his statures one can calculate probable body weights. My single step hopefully introduces fewer uncertain 89

AUSTRALOPITHECINEBODY SIZES procedures. Both McHenry and I have had to assume the same arm to leg ratio as in recent man and the evidence agaLnst this is still scanty. The usual interpretation of these (and other) postcranial reconstruc- tions has been that for the most part the larger half of them would be robustus and the smaller half africanus. The recognition here of male africanus*EiE-- female robustus stnttE-Eas ittTroduced Lwo new size categories. When these types are entered at the upper and lower ends, respectively, of the body size range, the size contrast between the two species becomes reversed. This reversal now accounts for the lrnpossible tooth to body ratio reported by Wolpoff (1973) for africanus. With africanus dentitions properly assigned to the larger average b;at srze they noE-dh6ilEout 17 rro2 oi grinaittg surface for every kg. of body weighE instead of the previously calculated 2I.2. The slightly largerlrobustug cheek teeth, combined with the much srnaller body sizes, now have 28 run' of grinding surface per kg. of body weight. These ratios are now entirely consistent with the higher degree of molarizaEion so ob- vious in the robustus teeth. In terms of Jo1ly's (1970) granivorous adapta- tions africanus noro fiEs his description of stage I and robustus takes this to the extreme of stage II. With the conventional body size allocations this simply wouldn't work, as robustus would have had relatively smaller teeth. Even if none of the involved calcul-ations above had been made, this reversal of body sizes would be required ln order to make relative toot,h sizes consistent. with the degree of molarization. This contrast in molarization between the two species might be related to the question of tool use. Oldowan sLone "toolst' occur in strata of australo- pithecine age and it ls still disputed who made them and for what use. The tool maker coul-d have been africanus or robustus, or both, or even neither if one assumes a higher hominid was al-ready there. From the evidence given here it would follow that africanus was the tool maker. Most of the evidence for a higher hominid can be dismTGeE--EJ bel-onging to young male africanus and female robus- tus. The difference in degree of molarization b"tr.6n E-Go species indicates 6-at for some time robustus has been selected for larger grinding teeth, while this was less true C-fficanus. Given the same kind of usage, the larger robustus teeth may be expectea to last longer (Wall-ace, 1973). -lt has been suggested that austral-opithecine tools were not so much used to extend the range of activities of the hominid biological equipment, but rather to prolong life as substitutes for worn out dentitions (Krantz, 1973c). Occlusal attrition in the australopithecines was so rapid that dental deaths would have occurred in what we regard as the prime of life, certain.ly less than twice the time it took to erupt all of the permanent teeth. Any method of extending effective tooth life would have been selected for. Increasing crordn surfaces is one method; adopting stone tools, especially in later years of l-ife, would be another. The division into two Australopithecus species rnay reflect the devel-opment of these two solutions. The lesser degree of molarization in africanus indicates they had some other method of extending dental 1-ife, presomiElfTEE-stone tools. This was a sLable biological solu- tion rrhich persisted without noticeable change for some two million years. This does not have to involve "culturet' any more than sea otters who also use sLone tools regularly. So africanus may well be our ancestor but they were not necessarily "human" in any realistic sense of the word. The average body size of the two sexes of africanus that is calculated here, 52 kg., is almost as large as genus Homo. The increase in brain size to H. erectus cannot be accounted for as part of an increase in body size. On the ottrer irana, the amount of brain increase is not as great as had recently been thought. The present scanty data would give a little over 500 cc. to South 90

G. S. KRANTZ

African africanus if the two sexes are weighed equal1y, aod robustus rates about 450 cc.

The Sex Ratios

Both species show atypieal sex ratios i-n the numbers of recovered and identified specimens. Combining all data given here on cranial and dental mater- ia1 with the tentative l-dentifications of the postcranials gives the following totals. For africanus we have 19 rnales and 26 females; for robustus we have 23 males and 17 females. With these absolute numbers the proportions of about three to tl,.7oin each species is distinctly abnormal . Since special attention was paid to identj.fying the two rarer categories, thelr actual numbers have probably been exaggerated in this compilation. The real sex ratio in each speei-es may be more like two to one when all specimens are correctly identified. One peculiarity about these ratios is that it is not the same sex that is rare in each species. The only evident regularity is Lhat the largest and smallest of the body sizes are the ones whlch are underrepresented. This strongly suggests carnivore selection. Leopards were most likely the rnajor predators working on these austra- lopithecines just as they are on baboons today. Brainfs (1970) reconstruction of the events around the South African caves indicates they were filled largely with bone debris from leopard ki11s falling into them from the trees around their operiings. Like any other predator, these cats tend to have a size pre- ference in their kills, even though they can and do on occasion bring down game of almost any size. A slight emphasis on prey weighing between 30 and 60 kg. would produce the observed results. Such a preference wouLd mean the leopards would often pass up a female robustus to take a larger victim, and would some- times avoid a male africanus because it was large enough to put up too much of a struggle. Of course, the younger mal-e africanus would be in the right sLze tange and not much trouble. The rarer extreme size categories must have actually occurred in normal- numbers ln 1-1fe, they had to die sometime, and their remains had to end up sonewhere. If these tended to be taken more conmonly by other carnivores they would often be deposited some place other than the leopardsr feeding trees at the caves. The "robustrr caves of Swartkrans and Kromdraai have a preponderance of rr,ale individuals (2L) over females (11) regardless of species. Conversely the "gracile" caves of Sterkfontein, Makapansgat, and Taung emphasize females (26) over males (11). This points to an additional size selection, presum- ably by the local carnivores, of somewhat larger quarry at the first two sites, and a smaller average size at the others. It is unlikely that most leopards preyed regularly on australopithe- cines just as they rarely take people today. This may be because hominids do not usually offer a rrst.riking platform" as do the quadrupeds. Leopard predation, then as now, could have been mainly the work of just a few defective individuals. In a single year, one such feline could kill and eat a hundred australopithecines and leave their remains at just a few favored spots. The entire South African assemblage nay be the work of only a handful of leopards, each one of which had its own preferred australopithecine size, and its own dining trees. Other factors may also be responsible for this site selection such as the local availabil-ity of various sizes of other game species. In any case, the detailed provenience of australopithecine remains more likely re- presents variations in the work of carnivores than any natural distribution of the victims. 91

AUSTMLOPITHECINEBODY SIZES Phylo genetic Implications

This reorganization of some of the australopithecine fossi-ls allows a potential simplification of hominid phylogeny as compared with many recent proposals. Two lineages are indicated, both at the Australopi,thecus grade of organizatLon. A. robustus was a separate line by three million years ago and developed the trorTilla- 8Ei-tal design of mol-arizatlon to the greatest degree. A. africanus had a dentition and body size more like our ovm and evolved into Homo erectus about the same time as robustus disappears from the fossil record. The factors separatlng africanus from robustus likely included their contrasting solutions to the problem of excessive tooth wear. The f-ncreased emphasis on cheek tooth size put a limit on the body size that could be attained by robustus without a major design change. Any further increase ln stature, say of 1-02, would bring an autonatic increase of 2I% in chewing surfaces, but also an increase of 337. in the amount of chewing re- quired for the l-arger body. The relative size of the grinding battery must increase disproportionately just to stay even with the needs of the larger body. Larger bodies without correspondingly increased grinding teeth would have brought the inevitable dental deaths into ever younger ages. The size limit for this design was apparentl-y attained in the East African robustus. By contrast, the use of broken stones to cut and crush some of their food would enable africanus to avoid this problem to a certain degree. Some tooth wear may have been relieved throughout their lives, and especially the loss of a number of teeth in later years could have been compensated for by this method. Larger body sizes become a possibility with this technological adapta- tion. Copers Law of increasing body sizes carries with it Renschrs corollary that brain sizes also increase, although at a slower rate. Even if this selec- tion for size were unrelaLed to intellectual functions, these larger brains would automatically have given their possessors a greater mental time span. The subsequent devel-opment of persistence hunting (Krantz, 1968) would expectabl-y involve those australoplthecines who were the most preadapted to accomplish lt. These same africanus also had the sirnple stone tools which, with little modi- ficatlon, would serve to kill game and butcher the carcasses. Gigantopithecus is probably another hominid (Weidenreich, 1945, Woo, L962, Robinson, 1972, and Eckhardt, L973), so their dentitions should be com- pared wlth Australopithecus. Their dental measurements (frorn casts) give an australopitheci.ne ratio of second inclsor to first molar. Three specimens show 19.9%, 24.57,, and 29.L"/", for an average of 24.5% which is just midway between the figures for robustus and africanus. The Indian specimen does not include incisors, but from the available space it looks like a lower, more robustus-like, percentage. In its C-P-P series, Gigantopithecus has the "leveled" africanus pat- tern, but there also seems to be an overl-ay of considerable sexual dimorphism in the relative size of the cani-ne. In five out of six instances the fl-rst premolar is larger than the second; ln four out of five i-nstances the first premolar is also larger than the canine. This pattern is particularly remin- iscent of some africanus jaws from Makapansgat. The importance of all this is that there are no dental characteristics to associate Gigantopithecus r^tith A. robustus which has sometimes been assumed. Gigantopithecus also cannot be taken sinply as an early, gigantic form of A. africanus, although the resemblances are greater. The canine is higher crowned than in africanus when it is not fu1ly worn down. The technically bicuspid premolars retain even more of the semisectorial design in their sloping 92 G. S. KMNTZ

mesiolabial surfaces. If geological dating would permit, Gigantopithecus could morphologically serve as a possible ancestor of Australopithecus. If dating does not permit, Gigantopithecus could be a continuation of an earlier, and far smaller, version which was also the progenitor of Australopithecus. This presentation should not (and will not) be accepted as a full demon- strati,on of the assertions made. The major points should be recapitulated here. Two morphological types can be distinguished by the relative emphasis on the canine vs. the second premolar, and other evidence of two degrees of molarization. Examples of both dental species come f,rom each of the major South African caves--they are not site specific. There is a major sexual dirnorphism in each species, rivaling or exceeding that found in the goril1a. Male africanus and female robustus are at opposite ends of the size range and have 6een ignored or taken as examples of early genus Homo. Up to now, com- parisons have been mainly betr.reen male robustus and female africanus which are the most common types. A more reasonable assignment of tfr" po"t"tt"ial bones makes africanus substantially larger than robustus. Ratios of tooth to body size aliEE-iilE degree of molarizatlon on11if ttre usual concepts of body sizes are reversed. Some of the species allocations of individual specimens given here could be wrong. There are probably more specimens that could be positively classified with this new approach, especially among the many recenr discoveries in East Africa. It was enough of a nightmare to pick out and keep track of as many specimens. as I have without any acceptable guidelines to follow. Ideally, the original specimens should all be re-examined in this light by unprejudiced, competent investigators. Unfortunately, many of the people who have access to originals or good casts have already committed themselves to the concepts of site allocationr large robustus, and early genus Homo. Yet the opinions of such authorities would carry the most weight in straight- enj-ng out the present contradictory situation. Independent of my own work, something like this has already begun with the possibility now being considered that Makapansgat may contain both specles. With the addition of this rather bold presentation perhaps more work will be stirnulated.

Literature Cited

Aguirre, E. L970. Identificacion de "Paranthropus" en Makapansgat. Cronical del XI Congreso Nacional de Argueologia, Merida, L969, pp.98-124. Bo1k, L. 1916. Problems of human dentition. American Journal of Anatomy 19 : 91-148. Brain, C. K. 1970. New finds at the Swartkrans australopithecine site. Nature 225:LII2-1119. Broom, R. and J. T. Robinson, 1949. A new type of fossil man. Nature 164: 322-23. Clarke, R. J., F. C. Howell, and C. K. Brain. 1970. More evidence of an advanced hominid at Swartkrans. Nature 225:L279'L222. Clarke, R. J. and F. C. Howell. 1972. Affinities of the Swartkrans 847 hominid cranium. American Journal of Physical Anthropology 372319-336. Coon, C. S. L962. The Origin of Races. New York: Knopf. Dart, R. A. 1955. Australopithecus prometheus and Telanthropus capensis. American Journal of Physical Anthropology 13:67-96. Day, M. H., R. E. F. Leakey, A. C. Walker, and B. A. Wood. L975. New hominids from East Rudolf, Kenya I. American Journal of Physical Anthropol ogy 42 : 461-47 6. Eckhardt, R. B. 1973. Gigantopithecus as a hominid. In (R. H. Tuttle, Ed.) 93

AUSTRALOPITHECINEBODY SIZES

Paleoanthropology, Morphology and PaleoecologY, PP. 105-129. The Hague: Mouton. (Published 1975). Holloway, R. L. L973. Early hornini{ endocast.s: volumes, morphology, and significance for hourinid evolutf.on. In (R. H. Tuttle, Ed.), Primate Functional Morphology and Evohlfuionr pp. 393-415. The Hague: Mouton. (Published 1975) . i Howe1l, F. C. L969. Paper read at'1969 Arnerican Anthropological Associatlon Meeting, New Orleans. Jolly, C. J. 1970. The seed eaters: a ne$t model of hominid dif ferentiation based on a baboon analogy. Man 5tI-26. Krantz, G. D. 1968. Brain size and huntlng abllity ln earllest man. Current Anthropology 9:450-51. L973a. The double descent of man. In (R. H. Tuttle, Ed.), Paleoan- thropology, Morphology and Paleoecologyr pp. 131-152. The Hague: Mouton. (Published 1975). 1973b. An explanation for the dlastema of Javan erectus skull IV. In (R. H. Tuttle, Ed.), Paleoanthropology, Morphology and Paleoecology' pp. 36L-372. The Hague: Mouton. (Publtshed 1975). 1973c. Romerrs rule and the origln of stone too1s. Man 8(4):631-33. I974, The earllest tool-maker from Kenya: Australopithecus or Homo? Newsletter of Lithic Technology 3 z6-9, Leakey, R. E. F. L973. Evidence for an advance Pllo-Pleistocene hominld from East Rudolf, Kenya. Nature 2422447-450. Lovejoy, C. E. L973. Bl-omechanical perspectives on the lower l-imb of early hornlnids. In (R. H. Tuttle, Ed.), Primate Functional Morphology and Evolutlonr pp. 29L-326. The Hague: Mouton. (Publlshed 1975). Lovejoy, C. O. and K. G. Heiple. 1970. A reconstruction of the fernrr of Austral-opithecus africanus. American Journal of Physical Anthropology 32:33-4O. Mann, A. E. L975. Paleodemographic Aspects of the South African Australo- pithectnes. Unlversity of Pennsylvania Publications in Anthropology No. 1, Philadephia. McHenry, H. M. L974. How large were the australopitheclnes? Amerlcan Journal of Physical Anthropology 40:329-340. L975, Biomechanical interpretation of the early hominid hip. Journal of HumanEvolution 4:343-355. Napier, J. R. L964. The evolution of bipedal walking in the homlnids. Archlves de Biologie 75 (Supplement), PP. 673-708. pilbearn, D. and S. J. Gould. 1974. Size and scaling ln human evolution. Science 185:892-901. Robl-nson, J. T. 1953. Telanthropus and its phylogenetlc signlficance. Ameri-can Journal of Physical Anthropology 11:445-501. L956. The dentition of the Australopithecinae. Transvaal Museum Memoir /19, Pretoria. 1960. The affinl-ties of the new Olduvai australopithecinae. Nature 185:456-58. L967. Variation and the taxonomy of the early horninlds. In (T. Dob- zhansky and M. K. Hecht, Eds.), Evolutionary Biology I:69-L00. New York: Meredith. I972. Early Horninid Posture and Locomotion. Chicago: University of Chicago Press. Tobias, P. V. L967. Olduvai Gorge, Volume II: The Cranium and Maxillary Dentition of Australopithecus (Zinjanthropus) boisei. Carnbridge: Unlversity Press. 94 G. S. KRANTZ

Tobias, P. V. L97I. The Brain in Hominid Evolution. New York: Columbia University Press. L973. Brain evolution in the hominoidea. rn (R. H. Tuttle, Ed.), Primate Functional Morphology and Evolution, pp. 353-392. The Hague: Mouton. (Published 1975). von Koenigswald, G. H. R. 1973. Australopithecus, Meganthropus and Ramapithecus. Journal of HumanEvolution 22487-49I. Walker, A. \973. New Australopithecus femora from East Rudolf, Kenya. Journal of HumanEvolution 2:545-555. Wallace, J. A. 1973. Dietary adaptati-ons of Australopithecus and early Homo. In (R. H. Tuttle, Ed.), Paleoanthropology, t"toiphology and Paleoecology, pp. 204-223. The Hague: Mouron. (published 1975). weidenreich, F. L945. Giant early man from Java and south china. Anthropological Papers of the American Museumof Natural History 4O: 1-134. Wells' L. H. L969. Faunal subdivisions of the Quaternary in Southern Africa. South African Archaeological Bulletin 24:93-95. L973. Forward from Taung. Journal of HumanEvolution 2:563-66. t'Tel-anthropust' Wolpoff , M. H. 1968. and the singl-e species hypothesis. American Anthropologist 70 :477-493. L970. The evidence for rnultiple homlnid taxa at Swartkrans. American Anthropologlst 72 z57 6-607 . L97.L. Metric Trends in Homi.nld Dental Evolution. Studles ln Anthropology No. 2, Case l{estern Reserve University, Cleveland. L973. Posterior tooth size, body slze, and diet in South African gracil-e australopithecines. American Journal- of Physical Anthropol-ogy 39 2375-393. Woo, J. K. L962. The rnandlbl-es and dentition of Gigan_topithecus. Paleontologia Sinlca L46.L-94.